Performance of Patient-Reported Outcome Measures in a Large Pragmatic Trial of Home-Based Palliative Care (HomePal): Methodological and Practical Considerations for Embedded Patient-Centered Design
Richard A Mularski,Brian Mittman,Eric Haupt,Susan Wang,Sarah Scholle,Carmit McMullen,Maureen Henry,Ernest Shen,Huong Q Nguyen,HomePal Research Group,Huong Nguyen,Ari Padilla,Mayra Macias,Janet Lee,Thearis Osuji,Kathleen Estrada,Rebecca Biddle,Byron Batz,Jasamin Disney,Teresa Martinez,Rose Roxas,Peter Khang,Dan Huynh,Mary Machado,Gina Andres,Angel Vargas,Richard Mularski,Britta Torgrimson-Ojerio,Madeline Peyton,John Brandes,Emily Schield,Phyllis Ramey,Chris Carlson,Paula Edwards,Vicki Krepps,Jennifer Black,Erin Bruner,Mary Ann McBurnie,Ning Smith,Suzanne Gillespie,Kim Funkhouser,Morgan Fuoco,Dea Papajorgji-Taylor,Phil Crawford,Kelly Kirk,Joe Cerizo,Kimberly Stewart,Daniel Vaughn,Meagan Shaw,Katie Vaughn,Joanne Lynn,Lynn Reinke,Charles Anderson,Summer Austin-Bowden,David Baker,Bill Clark,Janet Corrigan,Jennie Chin Hansen,Keung Luke,Thomas Lee,Carol Levine,Kristine Maberry,Diane Meier,Carol Joy Phillips,June Simmons,Judy Thomas,Henry Werch,Patricia Ganz,Mary Naylor,Soo Borson,Kevin Cain
DOI: https://doi.org/10.1089/jpm.2021.0164
Abstract:Background: The research enterprise has embraced patient centeredness in embedded efficient pragmatic trials, but limited data exist on using patient-reported outcomes (PROs) collected as part of usual clinical care for research. Objectives: We sought to assess the performance of different assessment methods for obtaining PROs in a pragmatic cluster randomized trial (HomePal study) designed to compare two models of home-based palliative care (HBPC). Design: Descriptive analytics, comparative trends, and psychometric performance of PROs collected in the HomePal study; measures included Edmonton Symptom Assessment System (ESAS), PROMIS-10, and others administered at baseline, 1, and 6 months. Setting/Subjects: HomePal was conducted in the Southern California and Northwest Kaiser Permanente regions in the United States; subjects were patients receiving HBPC and their caregivers. Measurements: We specifically compared ESAS obtained by research staff with those obtained by clinical HBPC nurses at the time of HBPC enrollment. We also compared ESAS completed by patients versus if done or assisted by a caregiver (proxy). Results: We enrolled 3533 patients and had 2205 ESAS measurements that met the criteria for analysis at baseline and 1447 at the one-month follow-up assessment. Research staff-obtained ESAS at admission to HBPC was higher overall (indicating more symptoms) than the clinically collected measure whether symptoms were reported by patients (31.7 ± 15.4 vs. 26.0 ± 13.4) or by proxies (36.9 ± 15.6 vs. 26.5 ± 13.5). These differences persisted with follow-up ESAS measures. Conclusions: We identified significant variability in PRO responses between different surveyors and whether proxy interaction was needed suggesting complex issues around PRO measure performance for pragmatic embedded trials. ClinicalTrials.gov Identifier: NCT03694431.